@Override
public StoreFuncInterface createStoreFunc(POStore store) throws IOException {
Configuration conf = ConfigurationUtil.toConfiguration(pc.getProperties());
StoreFuncInterface storeFunc = store.getStoreFunc();
JobContext jc = HadoopShims.createJobContext(conf, new JobID());
OutputFormat<?, ?> outputFormat = storeFunc.getOutputFormat();
PigOutputFormat.setLocation(jc, store);
context = HadoopShims.createTaskAttemptContext(conf, HadoopShims.getNewTaskAttemptID());
PigOutputFormat.setLocation(context, store);
try {
outputFormat.checkOutputSpecs(jc);
} catch (InterruptedException e) {
throw new IOException(e);
}
try {
outputCommitter = outputFormat.getOutputCommitter(context);
outputCommitter.setupJob(jc);
outputCommitter.setupTask(context);
writer = outputFormat.getRecordWriter(context);
} catch (InterruptedException e) {
throw new IOException(e);
}
storeFunc.prepareToWrite(writer);
return storeFunc;
}
public void testBinary() throws IOException, InterruptedException {
Configuration conf = new Configuration();
Job job = new Job(conf);
Path outdir = new Path(System.getProperty("test.build.data", "/tmp"), "outseq");
Random r = new Random();
long seed = r.nextLong();
r.setSeed(seed);
FileOutputFormat.setOutputPath(job, outdir);
SequenceFileAsBinaryOutputFormat.setSequenceFileOutputKeyClass(job, IntWritable.class);
SequenceFileAsBinaryOutputFormat.setSequenceFileOutputValueClass(job, DoubleWritable.class);
SequenceFileAsBinaryOutputFormat.setCompressOutput(job, true);
SequenceFileAsBinaryOutputFormat.setOutputCompressionType(job, CompressionType.BLOCK);
BytesWritable bkey = new BytesWritable();
BytesWritable bval = new BytesWritable();
TaskAttemptContext context =
MapReduceTestUtil.createDummyMapTaskAttemptContext(job.getConfiguration());
OutputFormat<BytesWritable, BytesWritable> outputFormat =
new SequenceFileAsBinaryOutputFormat();
OutputCommitter committer = outputFormat.getOutputCommitter(context);
committer.setupJob(job);
RecordWriter<BytesWritable, BytesWritable> writer = outputFormat.getRecordWriter(context);
IntWritable iwritable = new IntWritable();
DoubleWritable dwritable = new DoubleWritable();
DataOutputBuffer outbuf = new DataOutputBuffer();
LOG.info("Creating data by SequenceFileAsBinaryOutputFormat");
try {
for (int i = 0; i < RECORDS; ++i) {
iwritable = new IntWritable(r.nextInt());
iwritable.write(outbuf);
bkey.set(outbuf.getData(), 0, outbuf.getLength());
outbuf.reset();
dwritable = new DoubleWritable(r.nextDouble());
dwritable.write(outbuf);
bval.set(outbuf.getData(), 0, outbuf.getLength());
outbuf.reset();
writer.write(bkey, bval);
}
} finally {
writer.close(context);
}
committer.commitTask(context);
committer.commitJob(job);
InputFormat<IntWritable, DoubleWritable> iformat =
new SequenceFileInputFormat<IntWritable, DoubleWritable>();
int count = 0;
r.setSeed(seed);
SequenceFileInputFormat.setInputPaths(job, outdir);
LOG.info("Reading data by SequenceFileInputFormat");
for (InputSplit split : iformat.getSplits(job)) {
RecordReader<IntWritable, DoubleWritable> reader = iformat.createRecordReader(split, context);
MapContext<IntWritable, DoubleWritable, BytesWritable, BytesWritable> mcontext =
new MapContextImpl<IntWritable, DoubleWritable, BytesWritable, BytesWritable>(
job.getConfiguration(),
context.getTaskAttemptID(),
reader,
null,
null,
MapReduceTestUtil.createDummyReporter(),
split);
reader.initialize(split, mcontext);
try {
int sourceInt;
double sourceDouble;
while (reader.nextKeyValue()) {
sourceInt = r.nextInt();
sourceDouble = r.nextDouble();
iwritable = reader.getCurrentKey();
dwritable = reader.getCurrentValue();
assertEquals(
"Keys don't match: " + "*" + iwritable.get() + ":" + sourceInt + "*",
sourceInt,
iwritable.get());
assertTrue(
"Vals don't match: " + "*" + dwritable.get() + ":" + sourceDouble + "*",
Double.compare(dwritable.get(), sourceDouble) == 0);
++count;
}
} finally {
reader.close();
}
}
assertEquals("Some records not found", RECORDS, count);
}
@SuppressWarnings("unchecked")
@Override
public void run() {
JobID jobId = profile.getJobID();
JobContext jContext = new JobContextImpl(conf, jobId);
org.apache.hadoop.mapreduce.OutputCommitter outputCommitter = null;
try {
outputCommitter = createOutputCommitter(conf.getUseNewMapper(), jobId, conf);
} catch (Exception e) {
LOG.info("Failed to createOutputCommitter", e);
return;
}
try {
TaskSplitMetaInfo[] taskSplitMetaInfos =
SplitMetaInfoReader.readSplitMetaInfo(jobId, localFs, conf, systemJobDir);
int numReduceTasks = job.getNumReduceTasks();
if (numReduceTasks > 1 || numReduceTasks < 0) {
// we only allow 0 or 1 reducer in local mode
numReduceTasks = 1;
job.setNumReduceTasks(1);
}
outputCommitter.setupJob(jContext);
status.setSetupProgress(1.0f);
Map<TaskAttemptID, MapOutputFile> mapOutputFiles =
Collections.synchronizedMap(new HashMap<TaskAttemptID, MapOutputFile>());
List<MapTaskRunnable> taskRunnables =
getMapTaskRunnables(taskSplitMetaInfos, jobId, mapOutputFiles);
ExecutorService mapService = createMapExecutor(taskRunnables.size());
// Start populating the executor with work units.
// They may begin running immediately (in other threads).
for (Runnable r : taskRunnables) {
mapService.submit(r);
}
try {
mapService.shutdown(); // Instructs queue to drain.
// Wait for tasks to finish; do not use a time-based timeout.
// (See http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6179024)
LOG.info("Waiting for map tasks");
mapService.awaitTermination(Long.MAX_VALUE, TimeUnit.NANOSECONDS);
} catch (InterruptedException ie) {
// Cancel all threads.
mapService.shutdownNow();
throw ie;
}
LOG.info("Map task executor complete.");
// After waiting for the map tasks to complete, if any of these
// have thrown an exception, rethrow it now in the main thread context.
for (MapTaskRunnable r : taskRunnables) {
if (r.storedException != null) {
throw new Exception(r.storedException);
}
}
TaskAttemptID reduceId = new TaskAttemptID(new TaskID(jobId, false, 0), 0);
try {
if (numReduceTasks > 0) {
ReduceTask reduce =
new ReduceTask(systemJobFile.toString(), reduceId, 0, mapIds.size(), 1);
reduce.setUser(UserGroupInformation.getCurrentUser().getShortUserName());
JobConf localConf = new JobConf(job);
localConf.set("mapreduce.jobtracker.address", "local");
TaskRunner.setupChildMapredLocalDirs(reduce, localConf);
// move map output to reduce input
for (int i = 0; i < mapIds.size(); i++) {
if (!this.isInterrupted()) {
TaskAttemptID mapId = mapIds.get(i);
Path mapOut = mapOutputFiles.get(mapId).getOutputFile();
MapOutputFile localOutputFile = new MapOutputFile();
localOutputFile.setConf(localConf);
Path reduceIn =
localOutputFile.getInputFileForWrite(
mapId.getTaskID(), localFs.getFileStatus(mapOut).getLen());
if (!localFs.mkdirs(reduceIn.getParent())) {
throw new IOException(
"Mkdirs failed to create " + reduceIn.getParent().toString());
}
if (!localFs.rename(mapOut, reduceIn))
throw new IOException("Couldn't rename " + mapOut);
} else {
throw new InterruptedException();
}
}
if (!this.isInterrupted()) {
reduce.setJobFile(localJobFile.toString());
localConf.setUser(reduce.getUser());
reduce.localizeConfiguration(localConf);
reduce.setConf(localConf);
reduce_tasks += 1;
myMetrics.launchReduce(reduce.getTaskID());
reduce.run(localConf, this);
myMetrics.completeReduce(reduce.getTaskID());
reduce_tasks -= 1;
} else {
throw new InterruptedException();
}
}
} finally {
for (MapOutputFile output : mapOutputFiles.values()) {
output.removeAll();
}
}
// delete the temporary directory in output directory
outputCommitter.commitJob(jContext);
status.setCleanupProgress(1.0f);
if (killed) {
this.status.setRunState(JobStatus.KILLED);
} else {
this.status.setRunState(JobStatus.SUCCEEDED);
}
JobEndNotifier.localRunnerNotification(job, status);
} catch (Throwable t) {
try {
outputCommitter.abortJob(jContext, org.apache.hadoop.mapreduce.JobStatus.State.FAILED);
} catch (IOException ioe) {
LOG.info("Error cleaning up job:" + id);
}
status.setCleanupProgress(1.0f);
if (killed) {
this.status.setRunState(JobStatus.KILLED);
} else {
this.status.setRunState(JobStatus.FAILED);
}
LOG.warn(id, t);
JobEndNotifier.localRunnerNotification(job, status);
} finally {
try {
fs.delete(systemJobFile.getParent(), true); // delete submit dir
localFs.delete(localJobFile, true); // delete local copy
// Cleanup distributed cache
taskDistributedCacheManager.release();
trackerDistributedCacheManager.purgeCache();
} catch (IOException e) {
LOG.warn("Error cleaning up " + id + ": " + e);
}
}
}